Fixing issue with Magma groups that are not Ids in Gap

[jenpaulhus]

This should fix #6322.

The local page does not error:
http://localhost:37777/Groups/Abstract/sub/474552.a.108.b1.a1
http://beta.lmfdb.org/Groups/Abstract/sub/474552.a.108.b1.a1

Also, if you go to the group page itself and scroll down to the subgroup profile and click on the LaTeX name for a subgroup pf order 4394, you'll see that in the knowl there is message saying that we can't display the group instead of an error message.
http://localhost:37777/Groups/Abstract/474552.a
http://beta.lmfdb.org/Groups/Abstract/474552.a

[roed314]

I changed the message to be more helpful. @jenpaulhus, if it looks good to you and tests pass feel free to merge.

[jenpaulhus]

So I do like your message much better, but when the search is returned, the "Name" column includes "links" to the 3 groups of order 4394. Except the links don't work because the groups aren't actually in gps_groups.

I think this is just an issue of stripping off the url in the subgroups search which may be a different PR, but wanted to put it here before we forget it.

[roed314]

There's another issue (we do have some groups of order 6561). I'm working on a fix.

[roed314]

There are 2326 groups in the database whose order would have been caught by missing_subs(). I've added a new function to hopefully check them. Compare http://localhost:37777/Groups/Abstract/4692.a.

@jenpaulhus I pushed my changes to your branch, and these are more substantial, so take a look and let me know what you think!

[jenpaulhus]

Looked through this and it looks good. Just for the record if we need it later, here is a group with subgroup of order 6561 and NOT in our DB: 13122.cd and here is group with subgroup of order 6561 in our DB: 13122.bu

I checked the product of 4 primes and that worked as well.

Can't check p^7 p>11 as there don't seem to be any gps/subgroups of those sizes.

Wondering if we should make the "Hash" in the knowl a knowl itself?

[roed314]

Making Hash a knowl sounds like a good idea, but we need to write that knowl first (hash values aren't really exposed yet in many places).